mahdiyehebrahimi commited on
Commit
39bdf83
·
verified ·
1 Parent(s): 705d7fe

Create utcner.py

Browse files
Files changed (1) hide show
  1. utcner.py +88 -0
utcner.py ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import csv
2
+
3
+ import datasets
4
+ from datasets.tasks import TextClassification
5
+
6
+
7
+ logger = datasets.logging.get_logger(__name__)
8
+
9
+
10
+ _CITATION = """Citation"""
11
+ _DESCRIPTION = """Description"""
12
+
13
+ _DOWNLOAD_URLS = {
14
+ "train": "https://huggingface.co/datasets/mahdiyehebrahimi/utcner/raw/main/utc_train_ner.csv",
15
+ "test": "https://huggingface.co/datasets/mahdiyehebrahimi/utcner/raw/main/utc_test_ner.csv",
16
+ }
17
+
18
+
19
+ class DatasetNameConfig(datasets.BuilderConfig):
20
+ def __init__(self, **kwargs):
21
+ super(DatasetNameConfig, self).__init__(**kwargs)
22
+
23
+
24
+ class DatasetName(datasets.GeneratorBasedBuilder):
25
+ BUILDER_CONFIGS = [
26
+ DatasetNameConfig(
27
+ name="utcner",
28
+ version=datasets.Version("1.1.1"),
29
+ description=_DESCRIPTION,
30
+ ),
31
+ ]
32
+
33
+ def _info(self):
34
+ text_column = "text"
35
+ label_column = "label"
36
+ # TODO PROVIDE THE LABELS HERE
37
+ label_names = ['UndergraduateRegistrationExceptions',
38
+ 'CentralAuthentication&Email',
39
+ 'Senior(Registration,Deletion,Leave)',
40
+ 'Senior(Professor,Seminar,Proposal,Defense)',
41
+ 'Admissionwithoutatest', 'Calculateandchargetheinternet',
42
+ 'OfficeAutomation', 'Ph.D.(Admission,Registration,Removal,Leave)',
43
+ 'Ph.D.(Comprehensive,Research1and2,Opportunity)', 'Yekta|Nikan']
44
+ return datasets.DatasetInfo(
45
+ description=_DESCRIPTION,
46
+ features=datasets.Features(
47
+ {text_column: datasets.Value("string"), label_column: datasets.features.ClassLabel(names=label_names)}
48
+ ),
49
+ homepage="https://huggingface.co/datasets/mahdiyehebrahimi/utcner",
50
+ citation=_CITATION,
51
+ task_templates=[TextClassification(text_column=text_column, label_column=label_column)],
52
+ )
53
+
54
+ def _split_generators(self, dl_manager):
55
+ """
56
+ Return SplitGenerators.
57
+ """
58
+ train_path = dl_manager.download_and_extract(_DOWNLOAD_URLS["train"])
59
+ test_path = dl_manager.download_and_extract(_DOWNLOAD_URLS["test"])
60
+
61
+ return [
62
+ datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": train_path}),
63
+ datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": test_path}),
64
+ ]
65
+
66
+ # TODO
67
+ def _generate_examples(self, filepath):
68
+ """
69
+ Per each file_path read the csv file and iterate it.
70
+ For each row yield a tuple of (id, {"text": ..., "label": ..., ...})
71
+ Each call to this method yields an output like below:
72
+ ```
73
+ (123, {"text": "I liked it", "label": "positive"})
74
+ ```
75
+ """
76
+ label2id = self.info.features[self.info.task_templates[0].label_column].str2int
77
+ logger.info("⏳ Generating examples from = %s", filepath)
78
+ with open(filepath, encoding="utf-8") as csv_file:
79
+ csv_reader = csv.reader(csv_file, quotechar='"', skipinitialspace=True)
80
+
81
+ # Uncomment below line to skip the first row if your csv file has a header row
82
+ next(csv_reader, None)
83
+
84
+ for id_, row in enumerate(csv_reader):
85
+ text, label = row
86
+ label = label2id(label)
87
+ # Optional preprocessing here
88
+ yield id_, {"text": text, "label": label}